Abstract

Restriction endonucleases are widely used in laboratory applications from recombinant DNA technology to diagnostics, but engineering of restriction enzymes by structure-guided design and in vivo directed evolution is at an early stage. Here, we report the use of an in vitro compartmentalization system for completely in vitro selection of restriction enzymes. Compartmentalization of a single gene in a rabbit reticulocyte in vitro transcription/translation system serves to isolate individually synthesized enzymes from each other. In each compartment, an active enzyme cleaves only its own encoding gene, whereas genes encoding inactive enzymes remain intact. Affinity selection of the cleaved DNA encoding active restriction endonucleases was accomplished by the use of streptavidin-immobilized beads and dUTP-biotin, which was efficiently incorporated into the cohesive end of the cleaved DNA using a DNA polymerase. We confirmed that genes encoding active restriction endonuclease FokI could be selected from a randomized library. This method overcomes the limitations of current in vivo technologies and should prove useful for rapid screening and evolution of novel restriction enzymes from diverse mutant libraries, as well as for studies of catalytic and evolutionary mechanisms of restriction enzymes.

Received April 30, 2004; Revised and Accepted June 17, 2004

INTRODUCTION

Restriction endonucleases that occur naturally in bacteria as host defense systems have been widely used in laboratory applications from recombinant DNA technology to polymorphism detection for diagnostics. Although more than 3500 restriction enzymes, including 240 distinct specificities, have been isolated from various bacteria, many sequence specificities have not yet been discovered ( 1 ). Creation of artificial restriction enzymes that recognize a desired or long DNA sequence is expected to be useful for various applications, including genome engineering. However, studies on design and directed evolution of restriction endonucleases have been limited so far ( 26 ).

Numerous other enzymes have been improved by directed evolution, which usually involves iterations of random mutagenesis followed by selection using living cells ( 7 , 8 ). However, the scope of this in vivo approach is rather limited because of the restriction of library sizes based on transformation efficiencies of cells and the bias of selected clones with toxicity toward the cells ( 9 , 10 ). As an alternative approach, Tawfik and Griffiths ( 11 ) developed an in vitro compartmentalization (IVC) system that allows completely in vitro evolution of enzymes. In this system, artificial cell-like compartments provide the linkage of genotype (DNA) and phenotype (protein) for directed protein evolution ( 8 , 10 , 11 ). So far, IVC has been used for the selection of enzymes such as methyltransferases ( 12 , 13 ), a bacterial phosphotriesterase ( 14 ), and Taq DNA polymerase ( 15 ). In this study, we have applied the IVC approach for in vitro selection of restriction endonucleases, using FokI as a model enzyme.

FokI is a type IIS restriction endonuclease isolated from Flavobacterium okeanokoites and recognizes the nonpalindromic nucleotide sequence 5′-GGATG(N) 9/13 -3′ ( 16 ). FokI comprises two separate domains: a DNA recognition domain that binds the recognition sequence specifically and a catalytic domain that cleaves DNA nonspecifically in the presence of Mg 2+ ( 17 ). It has been demonstrated that the FokI catalytic domain can be fused to zinc finger DNA-binding proteins to yield hybrid restriction enzymes with novel recognition specificities ( 1820 ). One of the important goals of the study of in vitro protein evolution is the improvement of such artificial enzymes so that they are efficient enough for practical use as laboratory reagents, such as naturally occurring restriction enzymes. Here, we report the use of IVC for completely in vitro selection of restriction enzymes. We confirmed that the method is useful for selecting active enzymes from a randomized mutant FokI library.

MATERIALS AND METHODS

DNA construction

The FokI gene (1740 bp) was amplified from the genomic DNA of F.okeanokoites (obtained from IFO; Institute for Fermentation, Osaka, Japan) by PCR using the primers FokIF and FokIR (all primer sequences used in this study are listed in Table 1 ). The PCR product was digested with BamHI and PstI, and then cloned into a pT7K vector comprising a T7 promoter, T7·tag sequence ( 21 ), and kanamycin resistance gene. The resulting plasmid pT7K-FokI was used for the construction of mutant FokI genes as follows. A plasmid pT7K-FokIΔ containing an inactive FokI mutant gene (containing a mutation of the S140 codon TCA to a stop codon TAA) was constructed from pT7K-FokI by use of a QuickChange XL site-directed mutagenesis kit (Stratagene) using primers FokmutF and FokmutR. Similarly, plasmids containing a D450E, D467E or K469R mutant gene were constructed using primers EDKmutF and EDKmutR, DEKmutF and DEKmutR or DDRmutF and DDRmutR, respectively. Template DNAs for in vitro transcription/translation of the wild-type and the mutant FokI were amplified from the above plasmids by PCR using primers T7F and ORIR2.

Table 1.

Oligonucleotide sequences

Oligo
 
Sequence
 
CATmutF GCTTATAGCGGAGGTTATAATCTGCC 
CATmut3R GGCAGATTATAACCTCCGCTATAAGCNNNAGTNNNCACGATCACACCATAGTCGATCGGAGATCCGACAGTATAAATTGCTCCNNNCGGTTTCCTTGATCCACCCAAAT 
DDRmutF CGGTGTGATCGTGGATACGAGAGCTTATAGCGGAGG 
DDRmutR CCTCCGCTATAAGCTCTCGTATCCACGATCACACCG 
DEKmutF GATTACGGTGTGATCGTCGAGACTAAAGCTTATAGCGGAGG 
DEKmutR CCTCCGCTATAAGCTTTAGTCTCGACGATCACACCGTAATC 
EDKmutF GGATCAAGGAAACCGGAGGGTGCAATTTATACTGCGG 
EDKmutR CCGACAGTATAAATTGCACCCTCCGGTTTCCTTGATCC 
FokI112F CGTTGGGCACATGCTTTAGGATTTATT 
FokI112R AATAAATCCTAAAGCATGTGCCCAACG 
FokI459F CCTATTGATTACGGTGTGATCGTG 
FokI459R CACGATCACACCGTAATCAATAGG 
FokIF CGCGGATCCATGGTTTCTAAAATAAGAACTTTCGGTTGG 
FokIR CAACTGCAGTTAAAAGTTTATCTCGCCGTTATTAAATTTCCG 
FokmutF GTTGGACTTGCTTACTCTAAATAAGGTCGACGGCAGCGCCATTG 
FokmutR CAATGGCGCTGCCGTCGACCTTATTTAGAGTAAGCAAGTCCAAC 
ORIR2 CAGCAGATTACGCGCAGAAAAAAAGG 
T7F CGGCATATGATCCCGCGAAATTAATACG 
T7kumaF AATTTCATAGCCGTCTCAGCAGCCAAC 
T7kumaR GTTGGCTGCTGAGACGGCTATGAAATT 
Oligo
 
Sequence
 
CATmutF GCTTATAGCGGAGGTTATAATCTGCC 
CATmut3R GGCAGATTATAACCTCCGCTATAAGCNNNAGTNNNCACGATCACACCATAGTCGATCGGAGATCCGACAGTATAAATTGCTCCNNNCGGTTTCCTTGATCCACCCAAAT 
DDRmutF CGGTGTGATCGTGGATACGAGAGCTTATAGCGGAGG 
DDRmutR CCTCCGCTATAAGCTCTCGTATCCACGATCACACCG 
DEKmutF GATTACGGTGTGATCGTCGAGACTAAAGCTTATAGCGGAGG 
DEKmutR CCTCCGCTATAAGCTTTAGTCTCGACGATCACACCGTAATC 
EDKmutF GGATCAAGGAAACCGGAGGGTGCAATTTATACTGCGG 
EDKmutR CCGACAGTATAAATTGCACCCTCCGGTTTCCTTGATCC 
FokI112F CGTTGGGCACATGCTTTAGGATTTATT 
FokI112R AATAAATCCTAAAGCATGTGCCCAACG 
FokI459F CCTATTGATTACGGTGTGATCGTG 
FokI459R CACGATCACACCGTAATCAATAGG 
FokIF CGCGGATCCATGGTTTCTAAAATAAGAACTTTCGGTTGG 
FokIR CAACTGCAGTTAAAAGTTTATCTCGCCGTTATTAAATTTCCG 
FokmutF GTTGGACTTGCTTACTCTAAATAAGGTCGACGGCAGCGCCATTG 
FokmutR CAATGGCGCTGCCGTCGACCTTATTTAGAGTAAGCAAGTCCAAC 
ORIR2 CAGCAGATTACGCGCAGAAAAAAAGG 
T7F CGGCATATGATCCCGCGAAATTAATACG 
T7kumaF AATTTCATAGCCGTCTCAGCAGCCAAC 
T7kumaR GTTGGCTGCTGAGACGGCTATGAAATT 

N = A, C, G or T.

Library preparation

A library of mutated FokI with randomized amino acids at residues 450, 467 and 469 was constructed as follows. The N- and C-terminal fragments of the FokI gene were amplified by PCR using two sets of primers T7F-FokI459R and FokI459F-ORIR2, respectively, and then cloned with a TOPO TA cloning kit (Invitrogen). From the resulting plasmids, two DNA fragments were re-amplified by PCR using two sets of primers T7F-CATmut3R and CATmutF-ORIR2. The PCR products were purified using a QIAquick PCR purification kit (Qiagen) and then reacted with DpnI (New England Biolabs) for digesting the template plasmids. The reaction products were further purified using a QIAquick gel extraction kit (Qiagen) after 1% agarose gel electrophoresis. The purified DNA fragments were mixed and the library of full-length mutated FokI genes was amplified by overlap-extension PCR with KOD-plus DNA polymerase (Toyobo) using primers T7F and ORIR2. The PCR program was 25 cycles of denaturation at 94°C for 15 s, annealing at 65°C for 30 s and extension at 68°C for 320 s.

In vitro compartmentalization

In vitro transcription/translation reactions in water-in-oil emulsions were performed essentially as described elsewhere ( 22 ) using the following modifications. The TNT quick-coupled transcription/translation system (Promega) based on a rabbit reticulocyte lysate was used as the water phase, and mineral oil (Nacalai tesque) containing 4% (v/v) Span85 (Nacalai tesque) and 1% (v/v) Tween-80 (Nacalai tesque) was used as the oil phase. The DNA concentration was 50 pM. The emulsions were incubated for 2 h at 30°C. To recover the water phase, the emulsions were spun at 2000 g and 200 μl of supernatant was mixed with 100 μl of quenching buffer (phosphate buffered saline with 100 mM EDTA) and 900 μl of water-saturated diethyl ether. The mixture was vortex-mixed, then centrifuged, and the ether phase was removed. The water phase was washed with ether, dried and further purified using a QIAquick PCR purification kit.

Biotinylation and affinity selection of cleaved DNA

Incorporation of dUTP-16-biotin (Roche) into cohesive ends of digested DNA fragments in the purified water phase was done with ΔTth DNA polymerase (Toyobo) ( 23 , 24 ) for 10 min at 75°C. The reactions were purified using a QIAquick PCR purification kit and mixed with streptavidin magnesphere paramagnetic particles (Promega) and buffer B (10 mM Tris–HCl, pH 7.9, 1 M NaCl, 1 mM EDTA) on a rotator for 90 min at 4°C. The streptavidin beads were washed twice with buffer B, twice with buffer W (25 mM Tris–HCl, pH 7.9, 1 M NaCl, 1 mM EDTA, 2.25 M guanidinium chloride), with buffer B and with PCR buffer, and then resuspended in 15 μl of PCR buffer. The DNA captured on the beads was amplified by PCR with KOD-plus DNA polymerase using primers T7F and T7kumaR (30 cycles at 94°C for 15 s, at 60°C for 30 s and at 68°C for 140 s). The PCR products containing promoter and open reading frame were re-assembled with 3′-untranslated region (3′-UTR) DNA (amplified from pT7K using primers T7kumaF and ORIR2) by overlap-extension PCR using primers T7F and ORIR2. The products were electrophoresed onto 1% agarose gel, purified with a QIAquick gel extraction kit and used as template DNAs for the next round of selection.

Cloning and characterization of selected mutants

The PCR products amplified from beads in each round of selection were re-amplified by PCR with KOD-plus DNA polymerase using primers FokI112F and T7kumaR (25 cycles at 94°C for 15 s, at 60°C for 30 s and at 68°C for 90 s). The PCR fragments containing the randomized portion of the FokI gene were cloned using a TOPO TA cloning kit and sequenced using an ABI PRISM 3100 genetic analyzer (Applied Biosystems).

For the activity assay of the selected clones, full-length FokI mutant genes were reconstructed from the partial fragment of each clone as follows. Two common fragments were amplified from pT7K-FokI using primers T7F and FokI112R and from pT7K using primers T7kumaF and ORIR2. The mutated region of each clone was amplified from each plasmid using primers FokI112F and T7kumaR. The three fragments were purified and then joined by overlap-extension PCR using primers T7F (labeled with fluorescein) and ORIR2. The resulting DNAs (2 nM final) were used as the template for the rabbit reticulocyte in vitro transcription/translation system. As the negative control, 10% heparin was added to the reaction mixture. Digestion of the fluorescence-labeled DNA with the translated enzymes was analyzed by SDS–PAGE using a Molecular Imager FX (Bio-Rad Laboratories).

RESULTS AND DISCUSSION

Strategy for in vitro selection of restriction enzymes

The scheme for in vitro selection of restriction endonucleases by IVC is shown in Figure 1 . IVC utilizing water-in-oil emulsions was developed by Tawfik and Griffiths to cage a single gene per micelle for linking the gene (genotype) and the gene product (phenotype), and was first applied for in vitro selection of methyltransferase M.HaeIII ( 11 ). In the M.HaeIII selection procedure, a DNA encoding an active enzyme in a biotinylated DNA library was methylated in a compartment and rendered resistant to cleavage by restriction endonuclease HaeIII; thus, the biotin-label at the end of the DNA remained and could be used for affinity selection of uncleaved DNA encoding active methyltransferase with streptavidin beads ( 11 ). In this study, on the other hand, affinity selection of cleaved DNA encoding active restriction endonucleases is required. Thus, we examined methods of introducing biotin into cleaved DNA in a library recovered from emulsions. Our early attempts at ligating a biotinylated DNA adaptor with the cleaved DNA were unsuccessful because of the low efficiency of the ligation reaction (data not shown). As described below, success was achieved by using dUTP-biotin, which was efficiently incorporated into the cohesive end of the cleaved DNA with a DNA polymerase ( Figure 1 ).

Figure 1.

Schematic representation of the in vitro selection of restriction enzymes. ( 1 ) A DNA library encoding mutant restriction enzymes is introduced into a rabbit reticulocyte in vitro transcription/translation system. ( 2 ) A single DNA molecule compartmentalized in a reversed micelle of water-in-oil emulsions is transcribed and translated in vitro . In a compartment, a translated enzyme with high activity cleaves its encoding DNA. In other compartments, genes that do not express active enzymes remain intact. ( 3 ) The mixture of cleaved and intact DNA is recovered from the emulsions. A biotinylated derivative of deoxyribonucleoside 5′-triphosphate (e.g. dUTP-biotin) is specifically incorporated into the cohesive end of the cleaved DNA by DNA polymerase. ( 4 ) The biotinylated DNAs encoding active restriction enzymes are affinity-selected with streptavidin-immobilized beads. ( 5 ) The selected DNA is amplified by PCR and used for ( 6 ) the next round of enrichment or ( 7 ) cloning and sequencing.

Figure 1.

Schematic representation of the in vitro selection of restriction enzymes. ( 1 ) A DNA library encoding mutant restriction enzymes is introduced into a rabbit reticulocyte in vitro transcription/translation system. ( 2 ) A single DNA molecule compartmentalized in a reversed micelle of water-in-oil emulsions is transcribed and translated in vitro . In a compartment, a translated enzyme with high activity cleaves its encoding DNA. In other compartments, genes that do not express active enzymes remain intact. ( 3 ) The mixture of cleaved and intact DNA is recovered from the emulsions. A biotinylated derivative of deoxyribonucleoside 5′-triphosphate (e.g. dUTP-biotin) is specifically incorporated into the cohesive end of the cleaved DNA by DNA polymerase. ( 4 ) The biotinylated DNAs encoding active restriction enzymes are affinity-selected with streptavidin-immobilized beads. ( 5 ) The selected DNA is amplified by PCR and used for ( 6 ) the next round of enrichment or ( 7 ) cloning and sequencing.

Although we have previously used wheat germ extract ( 22 , 25 ) as an in vitro transcription/translation system compartmentalized in emulsions, here we used a rabbit reticulocyte lysate system, because the ligation and polymerization reactions of the FokI-cleaved DNA recovered from the wheat germ reaction mixture were inhibited, probably due to endogenous factors such as exonucleases that eliminate the cohesive end of DNA (data not shown). DNA from the rabbit reticulocyte reaction mixture accepted the incorporation of dUTP-biotin and thus had retained its cohesive end.

Addition of the rabbit reticulocyte in vitro transcription/translation reaction mixture to stirred mineral oil gave a water-in-oil emulsion with a mean droplet diameter (by vol) of 3.5 μm as measured by laser diffraction (data not shown). Accordingly, an emulsion made from 50 μl of reaction mixture dispersed in 1 ml of oil contains 2 × 10 9 compartments, each of which contains a single gene when the DNA concentration is <70 pM. In this study, all IVC experiments were thus performed at 50 pM DNA.

Enrichment of FokI genes

First, we confirmed that genes encoding active FokI restriction endonuclease could be selected from at least a 10 4 -fold excess of genes encoding inactive mutants in a model experiment.

Figure 2A shows the DNA construct encoding the active FokI or an inactive FokIΔ. The N-terminal T7·tag (MASMTGGQQMGRGS) was used as an epitope tag in western blotting analysis. The 5′-UTR contains a T7 promoter. The FokI recognition and cleavage sites were present only within the 3′-UTR. As the length of the 3′-UTR of mRNA is known to affect the protein yield in a cell-free translation system ( 26 ), we investigated the yield of FokI in the rabbit reticulocyte transcription/translation system supplemented with DNA template using various lengths of 3′-UTR (∼1000, 2000 and 3000 bp). The DNA with 3000 bp of 3′-UTR ( Figure 2A ) gave the highest yield of FokI (data not shown), and thus this DNA construct was used in further experiments.

Figure 2.

Selections for restriction endonuclease FokI by using IVC. ( A ) DNA templates for in vitro expression of FokI. The 1740 bp FokI gene is fused with the N-terminal T7·tag coding sequence (42 bp). The 108 bp 5′-UTR fragment contains T7 promoter. The 3045 bp 3′-UTR fragment contains 17 FokI cleavage sites. A mutant FokIΔ gene contains three nucleotide mutations (boldface), yielding a stop codon (*) and a SalI cleavage site (box). ( B ) Enrichment of the active FokI gene in multiple rounds of selection. Reaction mixtures containing 1:1000 or 1:10 000 molar ratio of FokI:FokIΔ genes were emulsified. The DNA after each round of selection was digested with SalI and analyzed by agarose gel electrophoresis.

Figure 2.

Selections for restriction endonuclease FokI by using IVC. ( A ) DNA templates for in vitro expression of FokI. The 1740 bp FokI gene is fused with the N-terminal T7·tag coding sequence (42 bp). The 108 bp 5′-UTR fragment contains T7 promoter. The 3045 bp 3′-UTR fragment contains 17 FokI cleavage sites. A mutant FokIΔ gene contains three nucleotide mutations (boldface), yielding a stop codon (*) and a SalI cleavage site (box). ( B ) Enrichment of the active FokI gene in multiple rounds of selection. Reaction mixtures containing 1:1000 or 1:10 000 molar ratio of FokI:FokIΔ genes were emulsified. The DNA after each round of selection was digested with SalI and analyzed by agarose gel electrophoresis.

The wild-type FokI and the mutant FokIΔ genes were mixed in a ratio of 1:1000 or 1:10 000, and then subjected to several rounds of IVC selection procedure. As shown in Figure 2B , three rounds of selection of the 1:1000 ratio of FokI:FokIΔ genes and four rounds of selection of the 1:10 000 ratio each resulted in a roughly 1:1 final gene ratio, indicating an enrichment factor of ∼10-fold per round. The efficiency is not so high, but IVC selection can be performed at the rate of one round per day and thus our system using dUTP-biotin should be useful for rapid selection of active restriction enzymes from a mutant library.

In vitro selection of a randomized FokI library

Next, we applied the dUTP-biotin method for the selection a mutant FokI library. In the library, codons corresponding to three residues D450, D467 and K469 in the catalytic domain were simultaneously randomized by PCR-based cassette mutagenesis ( Figure 3 ). The three residues have been predicted to be catalytic important residues on the basis of the sequence similarity with EcoRI and EcoRV ( 27 ) and the structural similarity with BamHI ( 28 , 29 ). So far, mutational analysis by alanine scanning has been performed for Asp-450 and Asp-467 ( 27 ), but not for Lys-469. The selection of our library consisting of 8000 (= 20 3 ) mutants is expected to provide information as to whether any residue other than alanine could be functional. Moreover, if only the wild-type is selected from the library, the three residues would be definitively proved to be essential for catalytic activity. The content of wild-type DNA in this library is (2/64) 3 = 1/32 768, theoretically.

Figure 3.

Randomized target residues involved in catalysis. ( A ) The structure of the FokI–DNA complex ( 28 , 29 ). Target catalytic residues (Asp-450, Asp-467 and Lys-469) in FokI are in red. The cleavage domain (sky blue) moves toward the cleavage sites (red arrows) on the DNA (yellow). ( B ) The partial sequence of FokI including the target residues. In a library, the three amino acids were randomized with NNN codons (N = A, C, G or T) as shown in red. To distinguish a wild-type gene in the library from the original native gene, four silent mutations were also introduced in the library (red), yielding a PvuI cleavage site (box).

Figure 3.

Randomized target residues involved in catalysis. ( A ) The structure of the FokI–DNA complex ( 28 , 29 ). Target catalytic residues (Asp-450, Asp-467 and Lys-469) in FokI are in red. The cleavage domain (sky blue) moves toward the cleavage sites (red arrows) on the DNA (yellow). ( B ) The partial sequence of FokI including the target residues. In a library, the three amino acids were randomized with NNN codons (N = A, C, G or T) as shown in red. To distinguish a wild-type gene in the library from the original native gene, four silent mutations were also introduced in the library (red), yielding a PvuI cleavage site (box).

After six rounds of selection, the total activity of the library DNA at each round was analyzed. As shown in Figure 4 , the DNA digestion activity began to appear in round 4 and increased in each round of IVC selection ( Figure 4 ; Emulsified), whereas no activity was observed when selection was conducted with non-emulsified reactions in which the linkage of genotype to phenotype was not achieved ( Figure 4 ; Non-emulsified). Since the total yields of synthesized proteins estimated by western blotting were almost constant through all the rounds of selection (data not shown), the increasing activity was not derived from an increasing quantity of enzymes but would be derived from an increasing proportion of active enzymes.

Figure 4.

DNA digestion activity at each round of emulsified or non-emulsified selection. Fluorescently labeled DNA after each round of selection was transcribed and translated en masse and analyzed by SDS–PAGE using an imaging analyzer. A result for the wild-type DNA (wt) is also shown.

Figure 4.

DNA digestion activity at each round of emulsified or non-emulsified selection. Fluorescently labeled DNA after each round of selection was transcribed and translated en masse and analyzed by SDS–PAGE using an imaging analyzer. A result for the wild-type DNA (wt) is also shown.

Cloning and characterization of the selected FokI mutants

The library DNA from the 4th and 6th rounds of emulsified selection was cloned and randomly chosen clones were analyzed by DNA sequencing. As listed in Table 2 , 21 of 76 clones (6th) and 2 of 36 clones (4th) had Asp, Asp and Lys at positions 450, 467 and 469, respectively, as in wild-type FokI. The DNA sequences of all clones were distinguished from that of the original wild-type DNA by pre-designed silent mutations ( Figure 3B ), indicating that the selected wild-type sequence was not a contaminant, but was derived from the library. In the sequences of 55 (6th) and 34 (4th) non-wild-type clones, the frequencies of the 20 natural amino acids at the three positions seemed to be random and did not show any obvious bias, such as acidic or basic amino acids, as in wild-type FokI. Further, these clones did not reveal distinct activity ( Figure 5A ). These results suggest that the genes encoding FokI mutants other than the wild-type were randomly selected from the library.

Table 2.

Amino acid sequences of selected clones

Round
 
Number of clones
 
Amino acid residues
 

 

 
Activity
 
  450 467 469  
21 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 25 other sequences    
 − 
 − 
 − 
 − 
 − 
 20 other sequences    
Round
 
Number of clones
 
Amino acid residues
 

 

 
Activity
 
  450 467 469  
21 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 − 
 25 other sequences    
 − 
 − 
 − 
 − 
 − 
 20 other sequences    

The amino acid residues at positions 450, 467 and 469 are shown for cloned mutants isolated from the library after four or six rounds of selection. The number of clones containing each sequence is indicated. Five clones at round 4 and one clone at round 6 contain stop codons. The activity of each mutant is also shown in Figure 5A .

Figure 5.

DNA digestion activity of FokI mutants. ( A ) Activity of selected mutants listed in Table 2 . The sequence of three mutated amino acids at residues 450, 467 and 469 in each mutant is shown at the top of each lane. The DDK protein has the same amino acids as in native FokI, though the DNA sequence is different from that of the wild-type FokI (wt). Fluorescently labeled DNA encoding each mutant gene was translated and analyzed by SDS–PAGE with an imaging analyzer. Heparin was added to inhibit translation. ( B ) Activity of site-directed mutants D450E, D467E and K469R.

Figure 5.

DNA digestion activity of FokI mutants. ( A ) Activity of selected mutants listed in Table 2 . The sequence of three mutated amino acids at residues 450, 467 and 469 in each mutant is shown at the top of each lane. The DDK protein has the same amino acids as in native FokI, though the DNA sequence is different from that of the wild-type FokI (wt). Fluorescently labeled DNA encoding each mutant gene was translated and analyzed by SDS–PAGE with an imaging analyzer. Heparin was added to inhibit translation. ( B ) Activity of site-directed mutants D450E, D467E and K469R.

The proposed active site motif of the restriction enzyme ( 27 ) permits Glu as well as Asp at position 467. Further, mutations between amino acids with similar properties are often tolerated without complete loss of activity. However, FokI mutants D450E, D467E and K469R were not selected in this study. To test whether these mutants were functional or not, we constructed the mutants D450E, D467E and K469R by site-directed mutagenesis and analyzed their activity. As shown in Figure 5B , no DNA digestion with these mutants was observed in this assay. Because the reaction conditions were not optimized for the FokI activity (e.g. the Mg 2+ concentration in the rabbit reticulocyte reaction is 0.5 mM, whereas the optimum for FokI is 10 mM), the mutants analyzed in this study may possess weak activity, but no mutant having activity comparable with that of wild-type FokI was present in the 8000 mutants.

Applications of in vitro selection of restriction enzymes

In this study, completely in vitro selection of a mutant restriction enzyme library was accomplished by using IVC and dUTP-biotin for the first time. This method allows screening of a large library of 10 9 –10 10 molecules within one day per selection round, and overcomes the limitations of current technologies of in vivo directed evolution (the library sizes of 10 4 –10 6 and the speed of several days per round) ( 26 , 30 ). Thus, the in vitro system should prove useful for screening and evolution of novel restriction enzymes from diverse mutant libraries.

Although restriction enzymes that can be selected by the use of dUTP-biotin are limited to those that create cohesive ends containing one or more nucleotide dAs, other biotinylated nucleotides can also be incorporated by ΔTth DNA polymerase ( 23 ). Biotinylated dUTP was also used in directed evolution studies of novel DNA polymerases by means of phage display ( 31 , 32 ). If an improved DNA polymerase with a higher incorporation efficiency of dUTP-biotin becomes available, it may lead to a higher enrichment factor in our in vitro selection system.

One application of this system would be the fine-tuning of hybrid restriction enzymes based on the linkage of a zinc finger DNA-binding domain to the FokI DNA-cleavage domain ( 1820 ). In principle, designed zinc finger proteins capable of recognizing desired DNA sequences can be produced ( 33 , 34 ). Thus, hybrid restriction enzymes with desired specificity can also be obtained, though further improvement of the activity would be required ( 19 ). Such enzymes that recognize a sequence long enough to specify a unique site in a genome would be useful for gene mapping and gene targeting ( 35 ) on the genome.

The in vitro selection system may also be utilized to study mechanisms of natural evolution of restriction enzymes by testing evolutionary scenarios in the laboratories. For example, in vitro co-evolution of restriction endonucleases and methyltransferases can make it possible to clarify the evolutionary mechanism of restriction-modification systems.

We thank Masato Yonezawa, Toru Tsuji and other members of our laboratory for experimental advice and useful discussions. We also thank Teruhiko Matsubara for helping with the preparation of Figure 3A . This research was supported by the New Energy and Industrial Technology Development Organization (NEDO) and a Special Coordination Fund from the Ministry of Education, Science and Technology, Japan.

REFERENCES

1.
Roberts,R.J., Vincze,T., Posfai,J. and Macelis,D. (
2003
) REBASE: restriction enzymes and methyltransferases.
Nucleic Acids Res.
  ,
31
,
418
–420.
2.
Lanio,T., Jeltsch,A. and Pingoud,A. (
1998
) Towards the design of rare cutting restriction endonucleases: using directed evolution to generate variants of EcoRV differing in their substrate specificity by two orders of magnitude.
J. Mol. Biol.
  ,
283
,
59
–69.
3.
Samuelson,J.C. and Xu,S.Y. (
2002
) Directed evolution of restriction endonuclease BstYI to achieve increased substrate specificity.
J. Mol. Biol.
  ,
319
,
673
–683.
4.
Rimseliene,R., Maneliene,Z., Lubys,A. and Janulaitis,A. (
2003
) Engineering of restriction endonucleases: using methylation activity of the bifunctional endonuclease Eco57I to select the mutant with a novel sequence specificity.
J. Mol. Biol.
  ,
327
,
383
–391.
5.
Zhu,Z., Zhou,J., Friedman,A.M. and Xu,S.Y. (
2003
) Isolation of BsoBI restriction endonuclease variants with altered substrate specificity.
J. Mol. Biol.
  ,
330
,
359
–372.
6.
Zaremba,M., Urbanke,C., Halford,S.E. and Siksnys,V. (
2004
) Generation of the BfiI restriction endonuclease from the fusion of a DNA recognition domain to a non-specific nuclease from the phospholipase D superfamily.
J. Mol. Biol.
  ,
336
,
81
–92.
7.
Farinas,E.T., Bulter,T. and Arnold,F.H. (
2001
) Directed enzyme evolution.
Curr. Opin. Biotechnol.
  ,
12
,
545
–551.
8.
Doi,N. and Yanagawa,H. (
2001
) Genotype-phenotype linkage for directed evolution and screening of combinatorial protein libraries.
Comb. Chem. High Throughput Screen.
  ,
4
,
497
–509.
9.
Cohen,N., Abramov,S., Dror,Y. and Freeman,A. (
2001
) In vitro enzyme evolution: the screening challenge of isolating the one in a million.
Trends Biotechnol.
  ,
19
,
507
–510.
10.
Griffiths,A.D. and Tawfik,D.S. (
2000
) Man-made enzymes from design to in vitro compartmentalisation.
Curr. Opin. Biotechnol.
  ,
11
,
338
–353.
11.
Tawfik,D.S. and Griffiths,A.D. (
1998
) Man-made cell-like compartments for molecular evolution.
Nat. Biotechnol.
  ,
16
,
652
–656.
12.
Lee,Y.-F., Tawfik,D.S. and Griffiths,A.D. (
2002
) Investigating the target recognition of DNA cytosine-5 methyltransferase HhaI by library selection using in vitro compartmentalisation.
Nucleic Acids Res.
  ,
30
,
4937
–4944.
13.
Cohen,H.M., Tawfik,D.S. and Griffiths,A.D. (
2004
) Altering the sequence specificity of HaeIII methyltransferase by directed evolution using in vitro compartmentalization.
Protein Eng. Des. Sel.
  ,
17
,
3
–11.
14.
Griffiths,A.D. and Tawfik,D.S. (
2003
) Directed evolution of an extremely fast phosphotriesterase by in vitro compartmentalization.
EMBO J.
  ,
22
,
24
–35.
15.
Ghadessy,F.J., Ong,J.L. and Holliger,P. (
2001
) Directed evolution of polymerase function by compartmentalized self-replication.
Proc. Natl Acad. Sci. USA
  ,
98
,
4552
–4557.
16.
Sugisaki,H. and Kanazawa,S. (
1981
) New restriction endonucleases from Flavobacterium okeanokoites (FokI) and Micrococcus luteus (MluI).
Gene
  ,
16
,
73
–78.
17.
Li,L., Wu,L.P. and Chandrasegaran,S. (
1992
) Functional domain in FokI restriction endonuclease.
Proc. Natl Acad. Sci. USA
  ,
89
,
4275
–4279.
18.
Kim,Y.-G. and Chandrasegaran,S. (
1994
) Chimeric restriction endonuclease.
Proc. Natl Acad. Sci. USA
  ,
91
,
883
–887.
19.
Kim,Y.-G., Cha,J. and Chandrasegaran,S. (
1996
) Hybrid restriction enzymes: zinc finger fusions to FokI cleavage domain.
Proc. Natl Acad. Sci. USA
  ,
93
,
1156
–1160.
20.
Smith,J., Bibikova,M., Whitby,F.G., Reddy,A.R., Chandrasegaran,S. and Carroll,D. (
2000
) Requirements for double-strand cleavage by chimeric restriction enzymes with zinc finger DNA-recognition domains.
Nucleic Acids Res.
  ,
28
,
3361
–3369.
21.
Doi,N. and Yanagawa,H. (
1999
) STABLE: protein–DNA fusion system for screening of combinatorial protein libraries in vitro .
FEBS Lett.
  ,
457
,
227
–230.
22.
Yonezawa,M., Doi,N., Kawahashi,Y., Higashinakagawa,T. and Yanagawa,H. (
2003
) DNA display for in vitro selection of diverse peptide libraries.
Nucleic Acids Res.
  ,
31
,
e118
.
23.
Ikeda,K., Inoue,H., Oka,M., Kawakami,B. and Kawamura,Y. (
1995
) A non-radioactive DNA sequencing method using biotinylated dideoxynucleoside triphosphates and ΔTth DNA Polymerase.
DNA Res.
  ,
2
,
225
–227.
24.
Nakajima,N., Ionescu,P., Sato,Y., Hashimoto,M., Kuroita,T., Takahashi,H., Yoshikura,H. and Sata,T. (
2003
) In situ hybridization AT-tailing with catalyzed signal amplification for sensitive and specific in situ detection of human immunodeficiency virus-1 mRNA in formalin-fixed and paraffin-embedded tissues.
Am. J. Pathol.
  ,
162
,
381
–389.
25.
Yonezawa,M., Doi,N., Higashinakagawa,T. and Yanagawa,H. (
2004
) DNA display of biologically active proteins for in vitro protein selection.
J. Biochem.
  ,
135
,
285
–288.
26.
Sawasaki,T., Ogasawara,T., Morishita,R. and Endo,Y. (
2002
) A cell-free protein synthesis system for high-throughput proteomics.
Proc. Natl Acad. Sci. USA
  ,
99
,
14652
–14657.
27.
Waugh,D.S. and Sauer,R.T. (
1993
) Single amino acid substitutions uncouple the DNA binding and strand scission activities of FokI endonuclease.
Proc. Natl Acad. Sci. USA
  ,
90
,
9596
–9600.
28.
Wah,D.A., Hirsch,J.A., Dorner,L.F., Schildkraut,I. and Aggarwal,A.K. (
1997
) Structure of the multimodular endonuclease FokI bound to DNA.
Nature
  ,
388
,
97
–100.
29.
Wah,D.A., Bitinaite,J., Schildkraut,I. and Aggarwal,A.K. (
1998
) Structure of FokI has implications for DNA cleavage.
Proc. Natl Acad. Sci. USA
  ,
95
,
10564
–10569.
30.
Gruen,M., Chang,K., Serbanescu,I. and Liu,D.R. (
2002
) An in vivo selection system for homing endonuclease activity.
Nucleic Acids Res.
  ,
30
,
e29
.
31.
Xia,G., Chen,L., Sera,T., Fa,M., Schultz,P.G. and Romesberg,F.E. (
2002
) Directed evolution of novel polymerase activities: mutation of a DNA polymerase into an efficient RNA polymerase.
Proc. Natl Acad. Sci. USA
  ,
99
,
6597
–6602.
32.
Fa,M., Radeghieri,A., Henry,A.A. and Romesberg,F.E. (
2004
) Expanding the substrate repertoire of a DNA polymerase by directed evolution.
J. Am. Chem. Soc.
  ,
126
,
1748
–1754.
33.
Beerli,R.R. and Barbas,C.F.,III (
2002
) Engineering polydactyl zinc-finger transcription factors.
Nat. Biotechnol.
  ,
20
,
135
–141.
34.
Uil,T.G., Haisma,H.J. and Rots,M.G. (
2003
) Therapeutic modulation of endogenous gene function by agents with designed DNA-sequence specificities.
Nucleic Acids Res.
  ,
31
,
6064
–6078.
35.
Bibikova,M., Carroll,D., Segal,D.J., Trautman,J.K., Smith,J., Kim,Y.-G. and Chandrasegaran,S. (
2001
) Stimulation of homologous recombination through targeted cleavage by chimeric nucleases.
Mol. Cell. Biol.
  ,
21
,
289
–297.

Comments

0 Comments